Patents
Literature
37results about How to "Improve resolution accuracy" patented technology
Efficacy Topic
Property
Owner
Technical Advancement
Application Domain
Technology Topic
Technology Field Word
Patent Country/Region
Patent Type
Patent Status
Application Year
Inventor
Method and device for recognizing and analyzing PDF text and picture based on OCR and storage medium
PendingCN119672747AReduce false recognition rateImprove resolution accuracyCharacter and pattern recognitionBiological modelsEngineeringDocumentation
The invention discloses a method and device for recognizing and analyzing PDF texts and pictures based on OCR and a storage medium, and particularly relates to the field of text and picture processing.The method for recognizing and analyzing PDF texts and pictures based on OCR comprises the following steps that S100, PDF documents are loaded, page analysis is conducted, page elements are extracted, and page layout information is processed; and step S200, for the extracted picture or the PDF page generated by scanning, performing high-precision OCR recognition by using a deep learning model, and converting characters in the image into editable texts. According to the method, efficient and accurate analysis of complex texts and picture contents in the PDF document is realized, the technical scheme not only improves the analysis precision and efficiency and enhances the flexibility and expandability of the system, but also promotes the development of digital and intelligent application of the document, and meanwhile, by saving resources and time cost, the method and the system have good application prospects. And remarkable economic benefits are brought to enterprises and individual users.
Owner:LEXIANG DIGITAL CO LTD
Submarine pipeline vortex-induced vibration analysis method based on multiple Fourier characterization PINN
ActiveCN119415868Aovercome limitationsexact decompositionSustainable transportationBiological modelsFrequency spectrumAlgorithm
The invention discloses a submarine pipeline vortex-induced vibration analysis method based on multiple Fourier characterization PINN, and the method comprises the steps: constructing a physical information neural network, determining a loss function of an initial condition, a boundary condition and a partial differential equation (PDEs) according to a submarine pipeline vortex-induced vibration model, and adding an anchor point to correct the network, so as to improve the fitting capability of the network; a Fourier transform technology is adopted, the key frequency of vibration and the frequency value of the key frequency are accurately identified, a Fourier mapping # imgabs0 # function is introduced, and a multiple Fourier mapping network is constructed to carry out refined mapping decomposition and feature extraction on vibration signals; according to the number and the value of the key vibration frequency, the mapping times and the hyper-parameter are dynamically adjusted to overcome the limitation of the PINN in processing complex vibration signals; integrating the spatial-temporal characteristics through point-by-point multiplication, and outputting a vibration learning result; according to the method, the analysis precision is improved, the capability of processing various high and low frequency superposition vibrations is enhanced, and the problem of spectrum deviation during high-frequency information learning is effectively solved.
Owner:FUZHOU UNIV
Time-frequency analysis method for non-stationary signal
InactiveCN119474835AImprove resolution accuracyHigh precisionMachine part testingDamping functionSoftware engineering
The invention belongs to the technical field of time-frequency characteristic detection of non-stationary signals, and provides a time-frequency analysis method of non-stationary signals, which comprises the following steps: S1, designing a damping forced vibration function group, and according to a frequency range and frequency resolution precision required to be detected by a to-be-analyzed non-stationary signal, calculating a time-frequency characteristic of the to-be-analyzed non-stationary signal according to the frequency range and frequency resolution precision required to be detected by the to-be-analyzed non-stationary signal; determining the number of functions in the function group and the resonant frequency of each function; s2, converting the damped forced vibration function group into a difference equation for numerical calculation; according to the method, a group of damped forced vibration functions is designed, and the resonant frequency, the damping coefficient and the damping function of each function in the function group are set according to the actual demand of the to-be-analyzed signal, so that the time-frequency characteristics of the non-stationary signal can be accurately analyzed; in addition, the frequency resolution precision of the method is determined by the set number of functions, can be flexibly adjusted according to requirements, can be infinitely increased to improve the precision, and can also be reduced by only analyzing important frequency bands.
Owner:NANTONG UNIV
Cable insulation aging detection device
PendingCN119535133AImprove resolution accuracyImprove detection efficiencyTesting dielectric strengthOptically investigating flaws/contaminationFluorescenceElectric cables
Owner:HAIKOU SUB-BUREAU GUANGZHOU BUREAU EHV TRANSMISSION CO OF CHINA SOUTHERN POWER GRID CO +1
Redundant control system and method of rotor position angle of aviation permanent magnet motor based on AD2S1210
The present invention belongs to the field of aviation permanent magnet synchronous motor control system, and discloses an aviation permanent magnet motor rotor position angle redundant control system and method based on AD2S1210. The space vector control FOC algorithm unit generates six PWM waves to drive the three-phase bridge inverter module to turn on and off by collecting the armature current, rotor position angle and speed information of the permanent magnet synchronous motor using the speed and current double closed-loop control strategy, thereby controlling the permanent magnet synchronous motor to reach the target speed or torque; the AD2S1210 rotary transformer decoding chip is used to collect the rotor position angle information and speed information of the rotary transformer coaxially installed on the permanent magnet synchronous motor; and the rotor position angle information and speed information are decoded and sent to the position angle redundant control algorithm unit; the position angle redundant control algorithm unit is used to judge the received rotor position angle information and output the correct position angle signal to the space vector control FOC algorithm unit, and convert the received speed information and output it to the space vector control FOC algorithm unit.
Owner:SHAANXI AVIATION ELECTRICAL
Turbine blade flutter evaluation method adopting single-channel harmonic balance model
ActiveCN121902531AReduce computing timeImprove resolution accuracySustainable transportationBlade accessoriesImpellerHarmonic balance analysis
The invention discloses a turbine blade flutter evaluation method adopting a single-channel harmonic balance model. The method comprises the following steps: 1, carrying out structural grid division on a turbine blade to obtain a vibration mode, a vibration frequency and a vibration pitch diameter of a vibration blade; 2, performing fluid grid division on the turbine blade to obtain grid coordinates under a cylindrical coordinate system; 3, aligning the coordinates of the structural grid and the fluid grid to obtain a vibration mode of the fluid grid; 4, adopting a Doppler frequency shift effect to obtain blade vibration frequencies, vibration pitch diameters and inter-blade phase angles of other blade rows except the blade row where the vibration blades are located; 5, realizing mutual transmission of upstream information and downstream information of the rotor-stator interface; 6, single-channel unsteady harmonic balance analysis is achieved; and step 7, evaluating the stability of the turbine blade flutter system. According to the method, the flutter of the multiple rows of turbine blades is evaluated, and the method has the characteristics of low calculation cost and high evaluation accuracy.
Owner:XI AN JIAOTONG UNIV
Solar flare whole-process observation and analysis method and system
ActiveCN121881164AReduce missed detection rateReduce overall process delayWeather condition predictionAlarmsNerve networkNetwork model
The invention provides a solar flare whole-process observation and analysis method and system, and the method comprises the steps: inputting an obtained coding grayscale image and an interference coding mask into a target spectrum neural network model, and obtaining a plurality of groups of reconstructed three-dimensional hyperspectral data; stacking a plurality of groups of three-dimensional hyperspectral data under the same optical filter wavelength according to the polarization dimension to obtain a plurality of groups of hyperspectral polarization data; performing multi-frame superposition noise reduction processing and space-time registration on the multiple groups of hyperspectral polarization data to obtain processed hyperspectral polarization data; performing dynamic inversion based on the processed hyperspectral polarization data to obtain a first physical parameter and a second physical parameter; and generating early warning information based on an evolution stage division result and an abnormal feature recognition result obtained through the first physical parameter and the second physical parameter. According to the method, the leak detection rate and the whole process delay of the key stage of flare observation are reduced, the analysis accuracy of the core physical parameters of the flare is improved, and the observation integrity, the real-time performance, the precision and the robustness are comprehensively improved.
Owner:TSINGHUA UNIVERSITY
A high-precision analysis method for silicon resonance frequency
ActiveCN119004082BHigh precisionImprove resolution accuracySustainable transportationNeural learning methodsAnalytic modelFast Fourier transform
The present invention provides a high-precision analysis method for silicon resonant frequency, and relates to the technical field of silicon resonant frequency analysis. The present invention obtains a theoretical frequency signal of a silicon resonator excitation power supply and an actual frequency signal of the excitation power supply, and performs fast Fourier transform to obtain a main frequency difference of a silicon resonator, extracts interference signal characteristics from the main frequency difference by establishing a main frequency analysis model, respectively obtains historical temperature, vibration, silicon resonant frequency data and interference signal characteristics of the silicon resonator, and forms a training data set of an ideal silicon resonant frequency under a standard environment, establishes a resonant frequency analysis model based on a neural network model, trains the model through the training data set, and further improves the analysis accuracy by constructing an optimization model of the analysis model, collects real-time interference characteristics of the power supply, temperature values and vibration values of the silicon resonator as inputs of the trained model, and outputs a high-precision silicon resonant analysis frequency.
Owner:HEFEI SHENGWEN INFORMATION TECH CO LTD
A spoken language understanding method, system and device based on cluster-guided incremental inference
This invention proposes a spoken language understanding method, system, and apparatus based on clustering-guided progressive reasoning. The method includes: acquiring a labeled sample set and performing semantic clustering to generate several semantic clusters; determining aligned semantic clusters for the target sentence and retrieving examples from each semantic cluster to construct an example set; performing intent detection and slot filling based on the example set to obtain a first intent prediction result and a first slot prediction result; guiding intent detection and slot filling based on the first slot prediction result to obtain a second intent prediction result and a second slot prediction result; and globally coordinating the first and second intent prediction results to obtain the final semantic parsing result. This scheme effectively achieves a balance between relevance and semantic diversity, while explicitly modeling the bidirectional dependency between intent detection and slot filling, realizing high-precision multi-intent spoken language understanding independent of training.
Owner:UNIV OF SCI & TECH BEIJING
AI-driven conversational ERP order generation method and system
The application relates to an AI-driven conversational ERP order generation method and system, and belongs to the field of resource planning systems. The method comprises the following steps: loading user permissions and collecting historical behavior sequences in response to a wake-up instruction; predicting an order intention and generating an active guidance prompt based on a time sequence coding model; extracting initial order elements by using a large language model, combining a static permission vector and a dynamic context vector for joint constraint optimization to generate accurate order elements; calling an ERP master data interface to obtain real business data after permission verification to generate a basic order data set; performing conflict resolution through a structured constraint rule library and a model output-rule bidirectional verification mechanism to generate a pre-order data object; and writing into a database to generate an official order after multi-modal fusion confirmation and risk control pre-checking. The application has the advantages that permission verification is pre-positioned, data is accurate and reliable, active guidance is provided, and multi-modal risk pre-checking is performed, so that order entry efficiency and compliance are significantly improved.
Owner:上海美武信息技术有限公司
A method for intelligent analysis and recommendation of scientific and technological innovation policies based on big data
PendingCN122285736AImprove resolution accuracyExcellent degree of personalizationRecommendation modelEngineering
This invention relates to the field of big data processing and intelligent recommendation technology, and discloses a method for intelligent analysis and recommendation of science and technology innovation policies based on big data. The method includes: format conversion and word segmentation of the original policy documents; inputting a structured policy dataset into a policy domain ontology model; inputting multi-source enterprise data into a multi-dimensional feature system and performing data preprocessing; performing semantic matching calculations on the policy knowledge graph and multi-dimensional enterprise profiles; analyzing using an enterprise policy preference model; adjusting the recommendation model parameters through a feedback data quality assessment model and online learning; analyzing using a cross-departmental policy association identification algorithm; and obtaining policy collaborative analysis results through policy association network graph construction and policy combination optimization algorithms. This invention, by constructing a multi-dimensional policy knowledge graph and a deep semantic understanding mechanism, can accurately analyze the complex logical relationships and fine-grained semantic elements in policy provisions.
Owner:HANGZHOU JIANGHONG TECHNOLOGY CO LTD
A laser reflection type pipeline leak detection method and apparatus
The application discloses a kind of laser reflection type pipeline leakage detection method and device, the detection method includes: by laser emitting unit, detection laser is emitted to pipeline inner wall;By photoelectric receiving unit, reflected laser after pipeline inner wall is reflected is received, and reflected laser is converted into voltage signal;Voltage signal is received by signal processing analysis unit, and whether pipeline is leaked is determined according to the amplitude of voltage signal.Such, by detection laser, pipeline inner wall is scanned, since the diameter of laser beam is small, the spatial resolution accuracy of measurable leakage point is high, and the detection of pipeline tiny leakage can be realized.
Owner:XIAN UNIV OF SCI & TECH
A method for analyzing the influencing factors of nearshore denitrification processes coupled with physicochemical factors.
ActiveCN121662223BOvercome the limitations of siled analysisovercome limitationsGeneral water supply conservationChemical processes analysis/designData setPrincipal component analysis
This invention relates to the field of environmental monitoring technology, specifically to a method for analyzing the influencing factors of nearshore denitrification processes coupled with physicochemical factors. The method includes: S1: collecting nearshore environmental samples to obtain a dataset of physicochemical factors; S2: calculating the coupling strength between each physicochemical factor based on the dataset, generating a coupling strength matrix; S3: using principal component analysis to reduce the dimensionality of the coupling strength matrix and determine a list of key influencing factors; S4: analyzing the contribution of each dominant factor to the denitrification rate based on the list of key influencing factors, generating contribution analysis results; S5: outputting an influencing factor analysis report on the nearshore denitrification process based on the contribution analysis results, including a ranking of key influencing factors and optimization suggestions. This invention provides intelligent, data-driven auxiliary decision support for nearshore ecological restoration and water quality improvement, and has good prospects for widespread application.
Owner:SECOND INST OF OCEANOGRAPHY MNR
Data query methods, devices, equipment, and media based on natural language recognition
ActiveCN121722904BImprove resolution accuracyOvercoming the problem of accurate mappingEngineeringResult set
This application provides a data query method, apparatus, device, and medium based on natural language recognition, belonging to the field of data query technology. The method includes: obtaining extended keywords based on user natural language queries; obtaining pattern information and business knowledge based on the extended keywords, user natural language queries, and a vector database; generating target prompt words based on the pattern information, business knowledge, user natural language queries, and system prompt word templates; generating multiple candidate query statements through a target large language model based on the target prompt words and model generation parameters, and determining the optimal query statement; executing the optimal query statement in a database test environment, obtaining an execution result set and capturing abnormal information; validating the execution result set and obtaining a verification result; if the abnormal information is empty and the verification result is qualified, then generating a visualization chart based on the extended keywords, the optimal query statement, and the qualified result set. This application can improve the accuracy of power data queries.
Owner:BEIJING JOIN BRIGHT DIGITAL POWER TECH CO LTD
Energy storage converter asymmetric fault analysis method based on virtual synchronous machine technology
The application discloses a kind of energy storage converter asymmetric fault analysis methods based on virtual synchronous machine technology, using the vector control strategy based on internal potential orientation, grid-connected point voltage is separated into positive and negative sequence;According to asymmetric fault scene adjustment control link, including using trap filter to filter out the two-frequency fluctuation component in inverter output power, the reference voltage in reactive-voltage loop is switched from rated voltage to grid-connected point positive sequence voltage, negative sequence is inhibited by balance current strategy;Based on reactive-voltage circuit equation and reactive-voltage control equation, establish the equation about internal potential;According to virtual impedance ring control equation, calculate current reference value before fault and substitute into the virtual impedance ring control equation after Laplace transform, establish the equation about fault current dq axis component;According to the equation about internal potential and the equation about fault current dq axis component, construct the analytical expression of fault current.The application considers the influence of inverter control link on fault current transient characteristic, and analytical precision is high.
Owner:NORTH CHINA ELECTRIC POWER UNIV +2
A method for intelligent automatic repair of three-dimensional scanning of paper documents
ActiveCN119027348Bimprove accuracyImprove resolution accuracyImage enhancementManufacturing computing systems
The present invention discloses a method for intelligent automatic repair of three-dimensional scanning of paper documents, and relates to the technical field of document data processing. The method comprises the following steps: collecting three-dimensional scanning parameters of paper documents; performing a first performance evaluation of three-dimensional scanning of paper documents and a second performance evaluation of three-dimensional scanning of paper documents according to the three-dimensional scanning parameters of paper documents; performing comparative analysis based on the three-dimensional scanning performance evaluation results of paper documents and a preset three-dimensional scanning performance threshold of paper documents; and adjusting the three-dimensional scanning of paper documents according to the comparative analysis results of the three-dimensional scanning performance of paper documents. The present invention improves the quality of data collected by three-dimensional scanning of paper documents by evaluating and analyzing the three-dimensional scanning parameters of paper documents and adjusting the three-dimensional scanning mode of paper documents, improves the accuracy for the subsequent generation of the final three-dimensional model of paper documents, and solves the problem of low quality of data collected by three-dimensional scanning of paper documents in the prior art.
Owner:CHINA NAT INST OF STANDARDIZATION
Child hospital BIM parameter optimization method based on multi-target GAN and LLM updating
PendingCN121980651AEnable accessibilityovercome deficienciesGeometric CADBiological modelsData setData source
A child hospital BIM parameter optimization method based on multi-target GAN and LLM updating comprises the following steps: step 1, extracting a newest standard from an authoritative data source by adopting LLM in combination with a GraphRAG technology, and generating a structured data set containing component parameters and corresponding standards thereof in combination with component data collected by a child hospital; 2, analyzing a component position relationship by using NLP, constructing a knowledge graph KG model, and storing the knowledge graph KG model in a Neo4j database to form a component-standard-position relationship network; 3, on the basis of the KG model, regularly verifying and updating a relation between a standard and a component through LLM + GraphRAG, and generating a dynamically updated optimization constraint condition; and 4, on the basis of the updated optimization constraint conditions in the step 3, through a multi-target GAN training generator and a discriminator, generating a core optimization parameter scheme, and outputting an optimized BIM design. According to the method, it is ensured that the optimization process adapts to the latest specification, and the defects of a traditional method in the aspects of multi-target balance, standard dynamics and engineering implementability are overcome.
Owner:XI'AN UNIVERSITY OF ARCHITECTURE AND TECHNOLOGY
Liquid crystal metasurface composite variable blazed angle grating and spectrometer
This invention belongs to the field of spectral analysis technology and relates to a liquid crystal metasurface composite blazed angle grating and a spectrometer. The liquid crystal metasurface composite blazed angle grating includes a planar substrate, a photo-alignment layer, and a liquid crystal layer stacked sequentially from bottom to top. The photo-alignment layer uses a photosensitive alignment material, and when irradiated by polarized light, its surface forms a molecular alignment pattern matching a preset composite phase distribution. The liquid crystal layer uses a high birefringence nematic liquid crystal material, and its internal liquid crystal molecules are arranged under the control of the molecular alignment pattern of the photo-alignment layer to form a liquid crystal metasurface structure matching a preset composite phase distribution. When composite light passes through the planar substrate and irradiates the liquid crystal metasurface structure, the liquid crystal molecules generate differentiated phase modulation of incident light of different wavelengths according to the gradient of the preset composite phase distribution, so that incident light of different wavelengths is diffracted and deflected at different angles, simultaneously realizing the dispersion and focusing imaging of composite light, thereby forming a spectrum.
Owner:SUZHOU UNIV
Quality identification method of Yuling paste based on ultrafast gas phase electronic nose technology
ActiveCN115856183BImprove nutritional anemia efficacyeasy to separateComponent separationMultivariate statisticalGas liquid chromatographic
The invention relates to a method for distinguishing the quality of Yuling paste based on an ultra-fast gas phase electronic nose technology, comprising: step one, homogenizing the Yuling paste after being treated with different steaming times to obtain a plurality of Yuling paste test samples; step two, using the ultra-fast gas phase electronic nose technology to perform odor analysis and detection on each Yuling paste test sample to obtain electronic nose gas chromatography information of each Yuling paste test sample, and obtaining an odor fingerprint of the Yuling paste according to all the electronic nose gas chromatography information; step three, using a PCA multivariate statistical method to screen differential chromatographic peaks in the odor fingerprint, performing qualitative material analysis on the differential chromatographic peaks to obtain a plurality of main differential volatile odor substances, selecting several of the main differential volatile odor substances as target characteristic differential volatile odor substances according to preset screening conditions, and determining an optimal steaming time range; and step four, establishing a discriminant factor analysis model DFA according to the electronic nose gas chromatography information.
Owner:BEIJING TONGRENTANG HEALTH PHARMA
A printed waste board collecting device
ActiveCN224463233UConvenient centralized transferImprove operational reliabilityMotor driveElectric machinery
The utility model discloses a kind of printing waste board collecting devices, it is related to printing waste board collection technical field, including printing waste board discernment mechanism, the printing waste board discernment mechanism includes support frame, the top of support frame is provided with printing plate conveyor belt, the top of support frame is fixedly connected with first mounting bracket;Printing waste board collection mechanism, the printing waste board collection mechanism includes motor, the surface of motor is fixedly connected with the surface of first mounting bracket, the one end of motor rotating shaft is fixedly connected with threaded rod.The utility model is formed by visual identification camera and lighting lamp and constitutes intelligent discernment system, can replace artificial and realize the accurate identification of waste board, avoid the missed detection or misjudgment caused by visual fatigue, substantially improve discernment accuracy;While motor drives threaded rod and drives plate lateral movement, cooperate with the suction cup of grabbing subassembly of cylinder and, can automatically complete the adsorption and transfer of waste board, completely abandon inefficient operation of artificial bending waist to take and place, significantly reduce manpower cost.
Owner:HUIZHOU HAOCAI PACKAGING PROD CO LTD
Instance-category feature association learning method for multi-person analysis
PendingCN119600651AImprove analysis efficiencyHigh precisionBiological modelsBiometric pattern recognitionPattern recognitionGraphics
The invention relates to an instance-category feature association learning method for multi-person analysis, and belongs to the field of computer vision and image graphics. The method comprises the steps of firstly obtaining human body multi-scale features; secondly, an instance-category feature representation module comprising a center head, an instance head and a category head is constructed, and an instance gravity center heat map, attitude features, instance features and category features are obtained respectively; then, an instance-category feature association module combining instance feature association learning and category feature association learning is constructed, instance-category feature association is carried out on the instance gravity center heat map, the attitude features, the instance features, the category features and the category kernel, and instance feature cosine and category feature cosine similar maps are obtained through learning; and finally, by calculating a direct index, performing instance-category feature fusion on the instance feature cosine similar graph and the category feature cosine similar graph, and finally generating a refined multi-person analysis result. According to the method, the multi-person image can be effectively analyzed, and the accuracy and efficiency of multi-person analysis are improved.
Owner:KUNMING UNIV OF SCI & TECH
Nuclear power DCS system timing method and system, electronic equipment and medium
PendingCN121806403AGuaranteed reliabilityImprove resolution accuracySynchronous motors for clocksSynchronisationTime errorTime deviation
The invention provides a nuclear power DCS system time calibration method and system, an electronic device and a medium, in the method, when a time calibration task is responded, whether a reference clock is forcibly calibrated or not is judged through local time and first calendar time of a dynamic clock, and the reference clock is forcibly calibrated without forcibly calibrating; the crystal oscillator error is quantized based on the second calendar time and the local time of the reference clock so as to capture the actual condition of the crystal oscillator deviation, then the dynamic clock timing rate is corrected by combining the first calendar time, the local time and the quantization parameter, and the obtained crystal oscillator deviation compensation parameter can dynamically adjust the dynamic clock timing rate instead of direct hopping timing. Time deviation caused by crystal oscillator errors is eliminated, the time reference of lower-level equipment can be effectively unified, and the problem of time errors caused by the crystal oscillator errors among the modules in the nuclear power DCS system is effectively solved.
Owner:CHINA TECHENERGY
Target recognition method, device and electronic equipment based on millimeter wave radar
ActiveCN119165463BImprove resolution accuracyReduce the risk of miscalculationWave based measurement systemsICT adaptationNoiseRadar detection
The embodiments of the present application provide a target recognition method, device and electronic device based on millimeter wave radar, which relate to the field of radar detection technology. The method comprises: generating a target range-Doppler spectrum according to a target echo signal; determining a first Doppler dimension noise floor and a second Doppler dimension noise floor according to the target range-Doppler spectrum, wherein the first Doppler dimension noise floor comprises: a Doppler dimension basic noise floor corresponding to a plurality of distance units, and the second Doppler dimension noise floor comprises: a Doppler dimension rainy day clutter area noise floor corresponding to a plurality of distance units; when it is determined that the weather mode of a target frame is a rainy day mode according to the first Doppler dimension noise floor and the second Doppler dimension noise floor, generating a constant false alarm rate detection result according to the target range-Doppler spectrum; and identifying the type of a target point to be detected according to the constant false alarm rate detection result, wherein the type of a target point to be detected comprises: a real target point and a false target point.
Owner:BEIJING SCI & TECH RUIXING ELECTRONIC TECH CO LTD +1
Touch display panel and touch module
PendingCN121957371AIncrease the number ofImprove resolution accuracyInput/output processes for data processingMechanical engineeringMaterials science
Owner:BOE TECHNOLOGY GROUP CO LTD
A Method and System for Unstructured Text Parsing and Question Answering Based on a Large Language Model
ActiveCN121009900Bimprove accuracyImprove resolution accuracyDatabase management systemsSemantic analysisTime informationData transformation
This application provides a method and system for unstructured text parsing and question answering based on a large language model. The method includes: acquiring raw unstructured text data from a data source at preset intervals; performing deep semantic analysis on the raw unstructured text data using a large language model to extract key entities and data structures; dynamically generating or updating an existing JSON structure template based on the key entities and data structures; converting the raw unstructured text data into standard structured data according to the JSON structure template; storing the structured data in a time-series database and establishing a multi-dimensional index; obtaining time information and query intent from natural language query information input by the user; and generating question-answering results or trend analysis results using a large language model based on the query intent and the structured data obtained from the time-series database according to the time information.
Owner:BEIJING YUNJIANXIN TECH CO LTD
An integrated intelligent terminal control system for customized image material sales
ActiveCN121481649BRealize quantitative evaluationRealize closed-loop controlMeasurement devicesDesign optimisation/simulation
This invention relates to the field of integrated image customization and intelligent sales decision-making technology, specifically to an integrated image customization material sales intelligent terminal control system. The system includes an intelligent terminal main control center for retrieving image data to be printed and user-selected material inventory unit information; an image reproduction complexity modeling unit for generating an image reproduction complexity index; a material medium interference potential analysis unit for generating a material medium interference potential; a process feasibility coupling calculation unit for calculating the process feasibility probability; and a dynamic sales decision reconstruction unit for comparing the process feasibility probability with a preset benchmark pass rate threshold. If the process feasibility probability is less than the preset benchmark pass rate threshold, a dynamic risk premium instruction is generated; otherwise, a basic selling price maintenance instruction is generated. This invention ensures that the control strategy can be dynamically optimized as equipment status ages or model changes, solving the problem that static models are difficult to adapt to dynamic performance changes in equipment.
Owner:BEIJING EGGPLANT BEAN NETWORK TECH CO LTD
Intelligent neuro-electric stimulation control method and system based on VTE system
PendingCN122251778Aaccurate captureImprove resolution accuracyMechanical/radiation/invasive therapiesPsychotechnic devicesTherapeutic effectElectrical stimulations
The application provides an intelligent neural electric stimulation control method and system based on a VTE system, relates to the technical field of intelligent control, and comprises the following steps: acquiring physiological signal data and stimulation state parameters, extracting features through a space-time correlation neural network to obtain a dynamic state vector; performing stimulation-response correlation mapping; generating an electric stimulation control scheme; and driving a stimulation device to perform an operation and collecting a feedback signal to dynamically update model feature weights. The application can realize personalized closed-loop stimulation control, improve stimulation precision and treatment effect, and reduce the risk of side effects.
Owner:HANGZHOU XIE TENG MEDICAL TECH CO LTD
Humanoid robot intelligent navigation method
PendingCN121995943AImprove understanding accuracyImprove resolution accuracyCharacter and pattern recognitionVehicle position/course/altitude controlHumanoid robot naoControl signal
The invention relates to the technical field of navigation, and particularly discloses a humanoid robot intelligent navigation method which comprises the following steps: step S100, receiving a multi-modal input instruction of a user, the multi-modal input instruction comprising a voice instruction or a text instruction; step S200, the multi-mode input instruction is processed, and a structured navigation instruction is generated; s300, acquiring environment information through an environment sensing module, and constructing a three-dimensional semantic map; step S400, based on the structured navigation instruction and the three-dimensional semantic map, planning a motion path of the humanoid robot through a path planning module, and generating a path trajectory; and S500, according to the path track, motion control is executed through a motion control module, the track is converted into a control signal of a joint motor of the humanoid robot, and the balance of the two feet is maintained in real time.
Owner:KUNLUN NUMBER (CHENGDU) TECHNOLOGY CO LTD
A method for analyzing time-varying control behavior of a driver based on hilbert-huang transform
ActiveCN119961599BControl behavior accuratelyAccurate analysis of control behavior
The application discloses a kind of based on hilbert-huang transform's analysis method of driver time-varying control behavior.Firstly, through improved empirical mode decomposition EMD and causal decomposition method based on EEMD, time domain signal is decomposed into a finite number of IMF, main causal mechanism in driver behavior is effectively extracted.Then, hilbert-huang transform is used to analyze the time-frequency characteristics of signal, and the control behavior of driver is understood in depth through causal analysis.Finally, using strong adaptive model identification method, based on the idea of two-step estimation method, the time-varying driver model parameters are identified, and the identification results are modeled precision evaluation.This method can not only accurately analyze the control behavior of driver, but also effectively process time-varying non-stationary signal, compared with wavelet, fourier transform and other methods, can reach high precision in time and frequency two dimensions simultaneously, help to improve flight safety.
Owner:BEIHANG UNIV
Body intelligent robot skill runtime data analysis and scheduling method
PendingCN122584290AMeet the needs of ultra-high precision scenariosImprove resolution accuracy
The present application relates to embodied intelligent robot technical field, specifically to embodied intelligent robot skill runtime data analysis and scheduling method. The method comprises: a skill data receiving module receives skill data packets of external data sources and completes format verification and integrity check; a multi-level semantic analysis engine carries out three-level progressive analysis of syntax layer, semantic layer and intention layer on the skill data packets, and generates executable action sequences; a real-time scheduling controller combines the current state of the robot and environmental information to prioritize and time slice allocate the action sequences, and generates a scheduling scheme; an adaptive optimization module continuously iterates scheduling strategy parameters based on historical execution data and real-time feedback. Through the technical framework of multi-level semantic analysis and dynamic scheduling cooperation, the present application solves the problems of information loss, response delay and insufficient adaptability in skill transfer process, and significantly improves the precision and task completion efficiency of robot skill execution in complex dynamic environment.
Owner:深圳复现范式科技有限公司